Serial analysis of gene expression (SAGE) is a highly quantitative method that determines the relative expression of each transcript present in any particular tissue of interest. We have recently performed a comprehensive analysis of gene expression using hierarchical clustering and our SAGE result from lung cancers. We found that the neoplastic state as well as the cancer subtypes could be distinguished based on gene expression patterns identified by SAGE. Furthermore, we identified 115 highly differentially expressed transcripts that could also provide the same distinction. Many of these genes were either previously identified as lung tumor markers or revealed molecular characteristics unique to the tumor tissue analyzed. cDNA array is a complementing method that allows for rapid analysis of many genes that are present at a higher level in the tissues. Using Affymetrix cDNA arrays, we have surveyed 50 primary lung tumors to determine the gene expression pattern among these tumors. Analysis are currently underway to determine the association of gene expression patterns obtained by cDNA array with the clinical and genetic phenotypes of lung cancer. Our array data also suggests that genes involved in G1/S and apoptosis pathways are frequently over expression in lung tumor tissues. During the past year, we have also constructed a lung tissue microarray that contains tumor tissues from 300 cases. We are expanding our study to include immunohistochemistry analysis to determine the interaction of proteins in different biological pathways and clinical outcome. We have also began to investigate the gene expression as well as genetic differences between the tumors in early and late onset lung cancer patients.